{"id":"W4403129964","doi":"10.1093/bjd/ljae377","title":"Do large language model chatbots perform better than established patient information resources in answering patient questions? A comparative study on melanoma","year":2024,"lang":"en","type":"article","venue":"British Journal of Dermatology","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":19,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Les Laboratories Pierre Fabre; Bristol-Myers Squibb Canada; Bristol-Myers Squibb","keywords":"Readability; Medicine; Computer science","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01716257,0.0003939686,0.0007591344,0.001010081,0.0003702272,0.001227361,0.0006471355,0.0006165619,0.002432599],"category_scores_gemma":[0.09143224,0.0002469626,0.0006918883,0.0006380271,0.0007291972,0.001776716,0.001280598,0.0006073109,0.0007742587],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00113565,"about_ca_system_score_gemma":0.0008136678,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003268622,"about_ca_topic_score_gemma":0.003459373,"domain_scores_codex":[0.9882648,0.00827353,0.0008740752,0.0009425641,0.001311193,0.0003338996],"domain_scores_gemma":[0.8924744,0.08740216,0.007321667,0.004055268,0.0062729,0.002473556],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.02645593,0.005669591,0.6120129,0.00399328,0.00135619,0.000851626,0.04559113,0.004969118,0.007278829,0.0004134111,0.004443626,0.2869644],"study_design_scores_gemma":[0.001224404,0.02352414,0.9266025,0.000588132,0.0009774275,0.001075545,0.01116609,0.02149653,0.005898595,0.0006320073,0.006626033,0.0001885089],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9977016,0.0002508015,0.0006360143,0.000116191,0.00001389245,0.0001357435,0.0002165879,0.00006077251,0.0008683193],"genre_scores_gemma":[0.9975061,0.0001390177,0.001202931,0.00009460891,0.00001887082,0.0001767618,0.0003554013,0.00002725316,0.0004790714],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01716257,"threshold_uncertainty_score":0.09076548,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04760364842800911,"score_gpt":0.3650601324547645,"score_spread":0.3174564840267554,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}